So I’ve been looking all over YouTube on how to model a head. I’ve came across one problem though. I was just curious on the quantity of polygons should be on a humans head in BGE. One thing I’ve observed from the tutorials is usage of subsurface modifier. So here’s the question…What the maximum amount of vertices should a player’s head have in BGE?
you can have as many polygons as there is no limit to them but the more you have, the more your game lags. If you have a basic scene you can have more polys but if you have a complex scene where every calculation counts, use less. Long story short, no more than 500 is about right.
Alright, thanks. I have a question about ears. Ears consume a large portion of polygons when in the head. If you make them low-poly, it will result in crappy quality. Is there an alternative?
In some benchmarks I did a fair while ago, when running on an i7 with GTX 970, BGE can do six render passes of a million polygons at 60FPS. This was with relatively simple shading and lighting. I can’t remember if it was the CPU or GPU that gave out at that point.
However, you can’t animate a million polygons at 60FPS. Nor should you expect physics to run well if you’re using triangle mesh collision at a million polygons.
Read from that what you like. It’s resource allocation, so that million polygons of rasterization can go into a single player with no level. Or you can do 900,000 poly level with a single 100,000 poly player.
Even on integrated graphics you can push a a million or so polygons with very simple shaders. On integrated graphics you’re likely to hit VRAM limits on your textures before you run out of GPU processing speed. (if you have 8 4096x4096 textures, you’ve gone past 512Mb of VRAM and many integrated GPU’s will stall).
These are relatively rough numbers, and were done about a year ago, so they may have changed now.
I was referring to a players’ head in Blender game engine. I assume you thought I was referring to Blender render?
the bge can handle millions of verts too. and the texture limitations still apply.
the real issue to think about is armatures. i assume your character will be rigged. the more verts the bones need to transform, the slower it will be. just how much slower i dont know.
They question you need to ask yourself is:
- How much detail will the audience see?
leading to a more precise question:
- how much detail will the audience see at what time?
Example:
- If you present a close-up (portrait), the details of the ears might be visible to the viewer
- if you show a body image, the details of the ear is so minor that it is hardly noticeable
- if the character has long hair, or an hat, the ears are not visible and you do not even need to model them.
- if you are presenting the components of an ear (e.g. for teaching) the details are highly important.
I think you might be asking the wrong questions. If you’re talking about a game, a genuine game, then you have far more to consider than the poly count for a characters head. That’s just one, arguably small element. And it’s dependent on what’s going on around the character in a given scene. It could be 500 polys or a 1000. Or possibly more.
What type of game are you aiming at? What’s likely to be going on in a heavy scene, where there’s plenty of activity and clutter? Is the world big and open? Is it clearly defined with a specific boundary? Will it have AI? What kind of texture resolutions would you favour?
The input of people with programming experience, specifically, will be valuable. But I think they’ll need more information about the game as a whole, if they’re to offer any advice.
It’s an open world game where you can pickup cloths. It’s going to be a multiplayer and both first and third person. It’s basically a survival game, not sure if that matters.
It matters, because this determines how much details are visible.
E.g. in a first person environment you do not need a player character at all, as it is not visible (you are looking away from it), unless you have mirrors.
In a multiplayer environment you need characters for the other players.
As you typically see them from distance, the max-details are determined by the min-distance to the other character.
When you want something like a wardrobe/inventory with character-preview you need to consider the details level there (which is independent from the detail level while running through the scenery.
When you want cinematic, it depends on the min-distance of the camera to the character. The difference to multiplayer is, that you know about the distance (as cinematic is typically not dynamic).
For all these different aspects you will need different levels of detail. When looking at them you should be able to determine the minimum distances to the camera and estimate the maximum of details you will need to model.
Be aware you should not show the max-details all the time, as it eats processing time, that you might need.